An evaluation of the accuracy of death certificate diagnosis of cancer for the state of North Carolina during 1975-1977 is conducted by matching such diagnosis with information gathered from hospital records. A major purpose of the study is to determine not simply whether cancer was present or absent at death, but to investigate the nature of the contribution of cancer to death, especially vis a vis other associated chronic illnesses (e.g., diabetes, heart disease, and stroke). Also the significant clinical events in the natural history of cancer caused deaths will be studied. The intent is to determine the completeness of death certificate diagnosis of major complications of the cancer underlying cause of death. In addition an analysis will be made of the appropriateness of the selection of cancer as the single underlying cause of death in the NCHS produced national mortality data. Analyses will examine the variation of the diagnostic accuracy and clinical significance of deaths reported as caused by cancer in national vital statistics data across a number of variables of interest (e.g., sex, race, age, cancer type, assocated chronic conditions, institution types and residence).